agree with Safetypee as above

However as the design functions consider such a balance with the pilots input, there doesn't appear to be linkage in the regulations between the pilot assumptions and their actual training syllabus. Often the initial training syllabus is predicated on historic , but not new design assumptions. An attempt is then made by the manufacturer to offer updated training for a new product, but even this falls outside of a well regulated requirement. We then end up with pilots who may meet standard training for an older design aircraft but are restricted in their understanding of a new product or a scare failure mode on an older product.

As an industry we are left with "best practices" which remain unregulated (cost and time prohibited)

The entire foundation of the problem is that the autopilot and autothrottles have become critical safety items for many of the world's pilots and airlines.

They should be nothing more than a tool to reduce pilot workload.

etudian

No, there is something seriously askew with the pilot who failed to monitor the developing upset. Automation is there to relieve the physical workload while doing the monitoring, not to relieve the thinking.

How does the pilot monitor the developing upset when the autopilot is sweating blood to avoid telling that fact?
Imho, Safetypee has it right, we've gradually evolved into a new world and the old rules don't really fit any more.

Split throttle levers, dramatically different engine indications left to right, yoke greatly displaced from the normal - all things that any pilot should be trained to notice (regardless of what they are flying).

Remember, we're talking a 737-500 - certified over 30 years ago at a time when the concept of "children of the magenta line' was just starting to form. Also a decade before Change Product Rule came into existence.
Pilots need to trained and capable of flying the model of aircraft they are going to fly - be it a 50 year old 737-200 or a brand spanking new, highly automated A320 NEO.

Quite often when a minor failure condition is allowed to migrate to a major failure by the pilot , there is no question of violating the regs over the design. This then leads to stronger warnings to increase the pilots knowledge.If this doesn't work then something like an idiot light is added in the cockpit. And even that doesn't work if the pilot never learns what that light is for and figured he doesn't need it when on automatics in a simple flight regime.

tdracer

Just wondering whether there is some way to alert the pilot that the autopilot is working outside of the normal parameters.
That way s/he gets a chance to actually look at stuff before it all goes haywire.
Right now, that transition seems to be abrupt, which is surely undesirable from an insurers perspective.

The autopilot will do its best until it reaches its limit, then it will trip out when it can’t cope anymore. If it could return to a normal state before tripping out, then it wouldn’t need to disconnect.

If I take control from the F/O, it’s likely to be because we’re in an unsatisfactory aircraft state which is my job to sort out.

Careful monitoring of the aircraft will give you a good idea of what to expect if you need to take control. If I see an increasing pitch attitude and decreasing airspeed while I’m PM, then I know my first action if I take over is likely to be stick forward.

etudiant

Actually looking at stuff and noticing it going haywire is the bare minimum expected of a pilot.

etudiant

Autopilot is out of my area of expertise, but as I noted in a previous post, Boeing did develop an autothrottle 'split' warning.
No idea why it wasn't implemented fleet wide - I've seen much 'lesser' fixes mandated by AD. Then again, given all the other ques that this crew failed to notice, I don't think it's a given that another que that something was wrong would have made much difference.
The real question remains - why didn't the crew notice? Were they really that asleep at the wheel, or was there something else going on that was distracting them (e.g. the pilot incapacitation that occurred on TAROM). Unless they find the CVR, we may never know.

Human performance is a spectrum; any training program necessarily makes a determination of the prerequisites to the training. In the middle of all that is the individual's awareness of their state in the universe. Then the individual gets to interact with operational realities, variables that involve anomalies, distractions, and pressures. Within that environment, when events go awry, observers get an opportunity to make commentary on the competency of the individual's performance. Since 1970, BAAA records indicate 11,170 commercial aircraft accidents where the aircraft was written off, where the system didn't work out well. Of those, fatal accidents are a subset. A percentage of these had no human causation at the operational end of the process, with design, manufacture, environmental and unlawful actions being the stated primary cause, arising from the industry state of the art. Yet, the human is also the means by which anomalies that were not or could not be foreseen are successfully surmounted.

Part 25 doesn't assume that the aircraft will go sideways undetected to the extent that the roll authority of the AFCS will be exceeded. Yet with a B74SP at 41,000' that occurs, and in that case, 3 professionally trained, qualified crew watch a plane decelerate and stall with power asymmetry, without rudder application, and during the ensuing gyrations, the crew's concern is that the problem is... an Attitude failure, while 3 ADI's 2 HSI's 2 RMIs all agree with the instruments. Strong plane.

Out of Jakarta, this aircraft had a minor anomaly that results in a LOC-I event that is not recovered from. The fundamental problem is not one of the training techniques for recovery (although that would be handy to have), the issue appears to be the failure to detect that an anomaly exists, which gets back to the failure of SA. If the crew does not recognize what has happened, then the recovery techniques are moot.

Yes, it is called scanning the instruments. In this particular case, the difference in N1 and deflection of control surfaces (which in the -500 is mirrored by column deflection) should be a fairly obvious sign that not all is going to plan. IF things are monitored properly, the pilots should be ready and alert by the time the A/P gives up and throws everything to their lap. Of course some are more capable of handling this than others, the rule of averages as had been discussed before.

Still awaiting for the CVR to be found and decoded/replayed.
However my money is on the CVR revealing that the crew were not, in essence, paying attention.

The fact that when the AP disconnected they didn’t cope must be a major cause for concern as to the quality of training & standards for the Regulator et al.

i-skew
 

andrasz, previously:-
‘… something seriously askew with the pilot who failed to monitor the developing upset‘
Concluding pilot ’failure’ without further justification adds nothing to a safety debate.
This event appears to been preceded by one or more flights which involved engine / auto-throttle malfunction, those crews managed the flight safely; how, why.

A more valued view of human performance contributions to safety is to consider why the previous crews succeeded in their monitoring, awareness, understanding, and acting, in apparently similar conditions.
A natural tendency is to start with the human; ok, but view the human as an asset.
Then consider the situation; similar conditions, but not necessarily the same situation.
Look for differences; perhaps an intermediate level off with higher thrust levels vs previous climb to cruise altitude. The crew interaction with the situation; distraction, focus of attention, ATC, …

This is normal operation, not seriously askew.
Insisting otherwise could indicate that our thinking and reasoning is serious askew, individual skew, ‘i-skew’.

For a systems approach to help understand this event, and as a human performance refresher, see in a recent publication sections 1.3 and 1.4 in https://www.easa.europa.eu/community...Regulators.pdf

P.S. ‘… some are more capable of handling this than others, the rule of averages as had been discussed before …‘ not the magenta line, but children of Lake Wobgone.

"RegulatorS et al" I would suggest. This is not the only crew in the last 1,5,10,20,30,40 years that has not arrived at the destination.

etudiant, re ‘… evolution, old aircraft, new world’
This is central to the safety debate; and as identified by lomapaseo, ‘As an industry we are left with "best practices" which remain unregulated (cost and time prohibited)’.

Historically safety intervention uses cost equivalence of fatalities vs action, placing a value on life when judging the cost-time tradeoff between modifying the aircraft or training the pilot.
The cost of modifying old aircraft might be disproportionately higher than newer models, also with larger fleets the cumulative cost could easily exceed training cost, which would benefit manufacturers - cost of modification (manufacturer pays) vs cost in training (operator pays).

The fallacy with training is the belief that its safety effect can be equated with modifying the aircraft, not so; ‘Training solutions are typically only assumed to be effective, not proven.’ *
Also, a cost-time evaluation must consider growing fleet size, future time in service, and thus greater exposure to the original risk. Which in turn requires more pilots to be trained, and many new pilots in following years - expensive, but hidden (forgotten) cost. How many aircraft / pilots since 737-500.
The industry's safety focus must change from past fatalities to future risk, particularly in an industry with a high level of safety. (Amalberti and Dekker argue that there will be little effective improvement above the overall 10^7 level).

In this 737 event, action based on fatalities could similarly misjudge safety tradeoff (older aircraft); and overlook the safety value in understanding preceding flights without fatalities. If action is warranted then perhaps dusting off the proposed mod; someone considered it to be of value at the time;- td

tdracer ‘The real question remains - why didn't the crew notice?’
I disagree, - why wasn't the proposed modification mandated. The safety thinking at that time needs to be understood. Also, why the industry apparently overlooked the findings of PSM+ICR, where the majority of issues in this discussion had been identified 20yrs ago. And irrespective of how many of the extensive training recommendations were accepted, why do we still continue to seek a training solution, which in failing to prevent this accident, implies that training has not been successful.

* PSM+ICR pages 67-68 https://www.faa.gov/aircraft/air_cer...ne_psm_icr.pdf

And for the inattentive: effective thrust monitoring requires comparing engine instruments (thrust parameter) not the input levers.

Monitoring the instruments is not that difficult, the FF, EGT, N2 will all give indications provided you are actually looking.. herein lies the problem, when the AI starts dancing and the crew have no capacity then the obvious gets ignored,,like the elephant in the room.
There are too many examples of Indonesian carriers trying to replace pilot skill with technology, look at the Lion Air crashes, all that happens is a more expensive piece of equipment ends up in the sea.
Human factors, early starts, complacency are all contributions but not valid reasons to crash a perfectly flyable machine.

And they clearly missed both.

lomapaseo

SLF here. I'm surprised and puzzled as to why a pilot would never learn the function of any indicator, instrument or system. Does training omit any of these?

I would hope that any new FO who discovered gaps in the syllabus would want to promptly fill them on his/her own time, for the sake of enhanced operational safety and efficiency. Plus, isn't there a natural desire to get to know the aircraft?

JM your assumptions would have been correct 30-40 years ago. With the advent of low cost carriers and the massive increase in air travel all over the globe then the standards of training and the motivation of individuals to know the aircraft inside out have dropped to the point that accidents of this type are not uncommon. Airlines don't want to spend anything above the minimum and a lot of them require the applicant to have done the type rating at their own expense. Aircraft manufacturers have dumbed down the manuals to the point that they only provide the information that they think you need to operate in normal operations. Regulators are quite happy to have pilots with a total of 200 hours experience operate as a First Officer. There is an art and science in the job of airline pilot, there is also an image and a status associated with being an airline pilot and there are too many pilots in the latter category who are not interested in the former.